266 Comparative PhOfWlogy. [APRIL,

a vigorous intellect and an earnt'st 80ul is contending for the Christian faith. Not that the work before us is a polemic againi:lt pantheism, or any heresy. It is the product of posi­tive and independent thought; its negative results are not, however, on that account, the less valuable.

That this delineation of so peculiar and original a course of thought has been, in all respects, successful, is too much to hope. If, however, it does not correctly express the leading features of the system examined, to those used to the pecu­liar phraseology and mode of thought of modern German philosophical writer!', it is not because a conscientious and painstaking endeavor has not been made.

ARTICLE II.

COMPARATIVE PHONOLOGY; OR THE PHONETIC SYSTEM OF THE INDO-EUROPEAN LANGUAGES.

BY BBIIJUUII W. DWIGHT, CLIIITOII, II. T.

[Continued from VoL XVI., p. 722.}

A Brief V"t.ew of tlte Sanskrit Consonants, in their relations to the Other Classical Languages.

THE different classes of consonants, in the Sanskrit, are all folluws:

(1) Gutturals. These are k, kh: g, gh, and n pronounced like our nasal n in ng and nk, as in sing and sink. Tbis nasal n is found only before gutturals: as in the middle of a word, or at the end of a word in place of m, if that word is succeeded immediately by one beginning with a guttural. K is represented, in Greek, by te, and in Latin by c (k) and q: as in Sansk. kapalas, the skull; Greek, teEtPa}..~; and Lat. caput. Kh is represented, in Greek, by X: 8S in Sanfik. nak­has, a nail j Gr. JvvE stem 8VVX (the 0 being euphonic); and

1860.] Comparative Phonology. 267

so khan, to dig, Gr. xalllEw, pure stem Xall. G is eq'liva­lent to the same sound in Greek and Latin: as in Sansk. sthag, to cover; Gr. rrrlryOJ; Lat. tego. Gh, as in Sansk. gharma; Gr. ":}EPp,O'>; Lat. formu!.!; Eng. warmth j is repre­sented by the aspirates of different organs in other languages. In the ease of laghu, light, it is represented, in Latin, by the labial v, in the word levis, light; while yet in the German leicht and English light, the original guttural form ill pre­servl"d.

(2) Palatals. These are ch, chh, j, jh, and n.1'his class of consonants may be viewed as derivative from the preced­ing, and but as a mere softened form of it. They occur only before vowels and weak consonant.s, as semivowels and na­Bals; while before strong contlonanbl they fall back at once ioto the class of gutturals from which they came. in the va­rious cognate languages, we find this class of letters repre­sented oftenest by gutturals j next, by labials, on aC<"ount of the mutual etymological sympathy so apparent in various languages between gutturals and labials; next in frequency, by some t-sound, as this is the initial element of the palatal sounds generally; and, last of all, by the sibilants. Thus compare

SANSKRIT.

panchan, five. pachami, I cook.

GREEK. , '7rEIITE

7rfUUOJ (for '7rfUuap.£ )

LATIN.

quinque. coquo (for coqua­mi).

jano, the knee. 'YOIIIJ genu. jam, to unite with. 'YalU'ill jribh, to open or draw apart.' 'Ypa.~E"', to scratch, to write.

ebh finds its equivalent, in Greek and Latin, in UIC and sc: as in <:hhnyll, a shadow, and UICt.<l; and also in chbinadmi, I cleave; and Lat. scindo (for scindami); chhauna, a cover­ing, and U"7JII~, a tent, as well as chhali and UIC{iM'>, the hit1c of an animal. When terminal in a root, it appears as g: as in Sansk. prachb, to ask; Lat. rogo for progo, Rtem prog (cr. also Lat. preeor, Eng. pray) ; and German, fragen.

(3) Linguals of a special sort, peculiar to the Sanskrit.

268 Comparative PhOtWlogy. [APRIL,

TheRe are written as t, th, 0, dh, n, each with a dot under­neath, to di8tinguish them from the ordinary dentals having the :mme symbols in their natural form.

(4) Dentals. This class embraces the common Iinguals of other languages, both simple and compound: aR d, dh, t, th, and n. D is sometimes interchanged with I in Greek and Latin: as in MICPVjJ4, a tear, and lacryma for dacryma; Sa~p (for SaF~p), a brot.her-in-law, and le"ir (Sansk. dcva­ras); and lingua, the tongue, archaic, dingua; and &1.4>"'1, a laurel, with its parallel form M4>"". Bopp regards simi­larly, and with good reason therefore, MP;Tra<; as repreSL'nting theS~nskrit dipa, a lamp, in a strengthened form; and so, -AtICO, in +jAt/Co<; he compares with drisa Prakrit disa, like. Compare also, in the same way, licet and SIfC'T], custom, right; an"d lorum, a thong, with oopa, a skin. The San8krit d, bet'ides being represented by its own simple equivalent in Greek and Latin, is, like dh, often represented by ~ (th) ; while dh itself, in addition to such an equivalent in Greek, is represented, also, by f and b in Latin. Thus compare:

SANSKRIT.

devas, a shining one. dvar, a door. duhitri, a daughter. dadhami, I place; stem, dha. madhu, an intoxicating drink. dhumas, smoke. tidhar, a teat.

GREEK.

~EO<;.

~vpa.

~lryaT"'p, n":tr'1!-U. stem, ~E. f'-~V. ~Vf'-O<;. ou~ap.

LATIN.

deus.

mel. fUIllUs.

uber.

Th, in Sanskrit, is never represented by ~ in the Greek, but always by 'T : as, in Sansk., stha, to stand, in the pres­ent, from tishthami, I stand, compared with 'hT'1f'-' for crla--rq­!-U (root, eTTa), Lat. sto, stare, stem, sta. So, compare Sansk. asthi, a bone, with OCTTEOV, Lat. os, stem, oss for ost; and also rat.h, a carriage, Lat. rota, a wheel.

(5) Labials. These are p, ph, b, bh, and m. Ph occurs rarely, while bh ill, like dh, of frequent occurrence. In Greek, cf>, and in Latin, f, represents, commonly, this aspirate. as in

1860.] Comparative Phonology. 269

Sansk. bhar, to bear, Gr. cf>Ep6>, Lat. fero; and also in bhu, to be, Gr. cf>.x", Lat. fui. In the Germanic languages, Sallsk. bb becomes also b, as in (ge)biiren, to bring forth, Eng.

" bear; compare, also, German fahren, to carry, Eng. ferry; and Germ. bin, I am, Eng. be, and Sansk. bhu.

In the dative plural ending -bu::!, Sansk. -bhyas, we see bh rE"presented by b, as its equivalent in Latin. In the in­terior of a word, indeed, tbe Latin prefers the medial labial (h) to the aspirates. Compare, in connection with Sansk. tubhyam, to thee, Lat. tibi; also, Sansk. abhi, both, Gr. I1p.­~,and Lat. ambo; and Sansk. nabhas, rabh and lubhyati with their Latin equivalents, nubes, a cloud, rabies, rage, and lubet or libet, it pleases. Sanskrit p, b, and m are each abundantly represented by their own simple equivalentl! in Greek and Latin.

(6) Semivowels. These are y, r, I, v. Y is, in Bound, our y, as in year. In Prakrit, as in Persian and Latin, it of­ten passes into j; as in Sansk. yuvan, young, Persian javAn, Lat. juvenis. In Greek, its equivalent is ~ Thus' compare yuj, to bind, and t'evryvVvtU and tcowVvtU, Lat. jungere, stem, jung and jug, as in jungo,andjugum; also, SaO!'lk.yava, barley, and t'ea for t'eFa j as al~o yas and t'e£v, to boil. So the ter­mination .at'co (for -at'a}£,) corresponds witb the similar San­skrit verbal ending -ay&mi. R is commonly represented by r in tbe other languages; and I sometimes passes over into r ill them: as in Sansk. lup and lump, to break, Lat. rum­po, perf. rupi. V has the sound of our English v, except after c01lsonants: as in tvam, where it is sounded like w. Neither v nor y can stand at the end of a word, since the voice cannot rest on them. As the semi vowels are of so flexible and flowing a nature, they easily interchange one with the otber, in tbe different languages, as not only an orig­inal I with r, as has been already indicated, but also an original n with 1. 'rhus compare Sansk. anyas, another, and Gr. aAA~ (for aMo~) and Lat. alius; and also Sansk. antaras and Lat. alter.

(7) Sibilants. Thcl:!e are C;. sh and B. H is also classi­fied here. The sibilant c; is very slightly aspirated. It ap-

23-

270 Comparative Pho7WWgy. [APRIL,

pears to have sprung from an original k; and, in Greek and Latin, k and c regularly correspond with it. The Gothic substitutes for it h, while the Lithuanian represenis it by a compound sibilant 8Z, pronounced like our ah. Thus Sansk. c;van, a dog, gen. c;unas, is, in Greek, ICtK"JI, gen. ICvvO~; in Gothic, hunds j and, in Lithuanian, szuo, gen. SZUI1~. SO, ac;vas, a horse, is, ill Latin, equus (pronounced, originally, as jf written ekus) and Lithuanian, uszwa. At the end of a word, and in the middle before strong consonants, it usually reverts to its original k-sound. With the tendency of this sibilant to vibrate between a hard and soft Bound, compare the double sound of c in our language, as 8 and k, or hard and soft j as, likewis£', in the French. In Italian, also, it has a double sound, as k and ch.

The sibilant ~h is pronounced as in English. Combined with k as in ksh, it is represented, in Greek, by ~, and in Latin by x: as, in Sansk. dakshina, Gr. &~u'k, Lat. dexter, Lith. deszine. It occurs sometimes initially and :Wine­times terminally: as in 8hash, six; where it is repn>8ent£'d, when initial, by 8 in Latin and the aspirate in Greek; and when terminal by x (0 in both languages; all in e, and Lat. sex; compare Lithuanian szeszi. At the end of a word, and in the middle before a strong consonant, as t and th, it passes into k and t, in Sanskrit. So, ill Greek /JICTW; Lat. octo; Italian otto; as compared with the Sansk. ashtan, eight; a similar style of interchanges appears ill the other languages. The sibilant. 8 is the ordinary s of other lan­guages. It is changed, in different cases, according to spe­cial euphonic rules, into C;, sh, r, and other letters, and only remains unaltered before t and tho

H was never admitted at the end of word:cl, or in the mid­dle before strong consonants. When coming into such posi­tions or conjunctiom .. , it passed, according to definite rules, into subdotted t or d, k or g j '\vhich it would be of no value to state or illustrate here, as they lie so exclusively within fhe bounds of specific Sanskrit scholarship, as such. The Sansk. h is repr~sented, often, by IC in Greek and c in Latin: as in Sansk. hard, hrid, and hridaya, the heart; Gr.

1860·1 Comparative Phonology. 271

ttapola. and JCfjp j Lat. cor, fltem cord, with which compare Gothic hairto j Germ. hertz j Eng. heart j and Lith. szirdis. In Greek, X is oft.en, also, lhe equivalent of the Sansk. h: as in Sansk. hima, Gr. Xf!lp.rJ)v, Lat. hiems j and also hrish, to rejoice, Gr. xatpt» j bansas, a goolle, Gr. X~v, Lat. anser for banser j and byas, yesterday, Gr. ~fS, Lat. beri for hesi j with be.4ernus, the adjective form of which, compare Germ. gestern and Eng. yesterday.

2dly. The Consonantal System of the several Classical Languages, viewed pathologically.

The true laws of consonantal combinat.ions, in reference to their proper euphonic effect, are better developed in Greek than in any, not to say all, other languages, besides the San­~krit. In no direction was their acute sense of the fitnes!! of things more exact and artistic j and in none was their skill more vigorously employed, than in their mode of constructing word-architecture, and adorning it according to their ideas of true taste. In the forms of words that they moulded and chillelled, or, in other language, in the additions, accommo­datiolls, abrasions, contractions, and prosodial changes, that they left "as the marks of their skill upon them, we see as in fixed type, the rules of art that they discovered and applied, in tbe mutual arrangement and harmonious difltribution of Bounds. Phonetic complications occur but on a very limited scale in Latin, whose laws of life and growth,in this part of its framework, are very simple.

That department of philology, which concerns itself with the affections or changes of letters and syllables, com~titutes the pathology of language, and embraces the whole range of mutilations and corruptions, whether effected by time, or dialectic causes, or the influences of climatic agency j l\s well as the whole range of euphonic additions, substitutions and suppression!', wrought by earnest determined hands, accnrding to real or supposed rules of art.

Letters once radical and characteristic of words in their original state, have dropped from their place, under the pres­sure of phonetic instincts and tastes upon them, like boughs encumbering the parent stem pf a tree, beneath the pruning

272 Comparative Phonolog-y. [APRIL,

knife; so that, in the scientific study of etymology, it be­comes often necessary to know, not only the course of the changes that have occurred, but also the laws that have determined their rise and progre.\!8.

The consonantal, like the vowel, elements of speech, have their different degrees of weight; and their weight is but another name for the amount of their phonetic force, or the densit.y, all it were, of their phonetic substance. The breathing h is lighter even than the vowels; to which the aspirate:! and semi-vowels stalld next in order; then follow the liquids and in the following sequence, from light to heavy, r, I, n, m. The heaviest of all the consonantal sounds are the mutes; and in the order for increasing weight of middle, smooth and rougb. So also labials and palatals of the same several classes, smooth, middle and rough, as p and Ie, band g, are heavier than the corresponding dentaltl of each class respectively, as t compared with p aud k, and 80 also d compared with band g. These 8ubtle mechanical relations of sounds to each other, indicate the directions in which the inward forces at "rork upon language, to modify its combinations, exert their energy.

As the facts and laws that pertain to consonantal combi­nations are intimately interwoven with those pertaining to consonant.al change!'l, they must, many of them, in order that either should be properly comprehended, be exhibited to­gether in one view.

(1) Generally: with a view of the general laws of change in word-forms. These laws of change are the following:

§ 1. The tendency is always, in the course of time, and in the passage of words from one country to another, for­ward:! from complicated to simple forms, and not backwards from simple to complicated. Time abrades and round:! off words in its perpetually flowing stream, as it does stones and boulders on the floor of the ever heaving sea.

§ 2. The greatest mutilations in the volume of words occur in their tenninal, rather than in t.heir initial syllables; although in the latter, changes of sillgle)etters occur mare frequently than in_the former.

1860.] Comparative Phonology. 273

§ 3. Vowels are much more senflit.ive to changcs in the volume of a word, and correspond more instinctively with them, than consonants.

§ 4. In vowel-changes the course of change is, for the most part, from the primary to tbe secondary vowels, and not backwards. The primary vowel, a, call be transformed into any of the other vowpls i but they do not revert to it. So ill Latin, e and 0 often settle down into the weaker vowels, i and u.

§ 5. The interchanges of· consonants with each other, which constitute a very large clas~ of all phonetic changes, are made on the following prindples:

1. Inasmuch as sounds made by different organs would, when proximate, often jar phonetically upon paeh other, or, which is the same thing, wuuld require ~pecial effort to be distinctly uttered in conjunction, they are harmonized on the principle that a smooth mute must precede a ::!mooth, a medial a medial, and a rough, a rough: as in ma and lfJooJU"', 8ICTCI) and 8ryBcm, ryp~, rypa7rTor; and rypaf3&qv. The law, stated in its simplest form, is this: consonants brought into immediate juxtaposition must be made homogeneous. Thus, ,,~ and X~ become ry8 i 181 and tf.!s become xJ!1; "(T and xr become ICT i 7r8 and tf>8 become fJ~; 7r~ alld ffJ become "'~ ; and f3T and <In" become 7rT.

2. Homogeneous consonants of different organs are often exchanged for each other.

(a) Semi-vowels and aspirates, one with the other; I\S

band 8, in ;, and sex, mci and scptem; and h and v, in ItrtrEpor; (FEtrtrEpor;) and vesper. So f in Latin bpcomes h in Spanish, all in Lat. filius, Span. hijo; and filum, Span. hilo.

(b) Different liquid::!, one with the other, as I and r, I and D, m and n; examples of which will be furn"ished hereafter, under the head of Substitution of Sounds for each other.

(c) Different mutes, one with the other, in eaeh of the three kindtl respectively, smooth, middle, and rough; abun­dant illustration::! of which will be furnished hereafter.

1 ID .he preposition lit, It remaiDl DuchaDged before I), IU in tfCt9ff1l1.

274 Comparative Pkonoloouy. [APRIL,

3. Homorganous consonant!', or those of anyone specific class, as labials, palatals or dentals, severally, may I't"adily pas1' into others of the same class, that is, others made by the same organs. The following are a few among many specimens: {30VMJUU and volo, I wish; X/iJl4 and hiems. wintry weather; XOfYTCK and hortus, a garden; uv (Aeol. TV) and tu, thou; ,dUCK and medius, middle; {3purrx (for JIf'OTfR, cr. Sanskrit marttas, Lat. mortuus, dead, from San~k. mri, to die, Lat. mori); and TlnrrOVU£, Aeol. and Dor. TV7rrOVT'

and TlnrrOJlU£.

The styles or forms of consonantal changes are varions, as A. Substitutions. B. Insertions and Additions. C. Suppressions. D. Weakened Consonantal Forms. E. Strengthened Consonantal Forms.

A. Substitution. Thi1' is of t.wo kinds: 1. Literal, or pertaining to a mere change of letters. 2. Topical, or pertaining to a change of place or order, in

respect either to a mere letter, or an entire syllable. 1. Literal Substitution. This is of two Idnds: (1) General, or weak. (2) Directly assimilative, or intensive. Assimilative Imbstitutioll occurs, when, by the strong

phonetic aUraction of anotht'r letter preceding or succeeding it, a consonant is changed to the same letter, or to one directly homogeneous with it; while, by general or weak subslitution is described any other change of a consonant., made under the influence of weak phonetic attraction, or or indeterminate subtle affinities of any kind, or for the mere sake of avoiding phonetic monotony.

A'S the modes and forms of substitution are so oftt'n the same in both Grcek and Latin, and these two languages are so cognate and correlated in every way, illustrations will be drawn indiscriminately from them both.

(1) General or weak substitutions occur in each of the diflerent classes of consonants.

I. Palatals or Gutturals. These are in Greek It, "I, X, and in Latin c, g, ch.

1860.] Comparative Phonology. '¥75

§ 1. The gutturals when followed by u become in Greek E, which, therefore, alwaY:3 repre~ents as a double consonant either /C, "I or X compounded with u. As in Latin g is ex­changed before sand t into c, x commonly rl'presents c + 8, but often also g + 8, and 80metimes v + s, as in vixi, perf. of vivo, for vivsi, and nix (gen. nivis) for nivs.

§ 2. The gutturals, when originally followed by " were afterwards changed to u or T; and the vowel was itself also 8ubi:lequently assimilatt'd to the same lett.er, which thus becom~ double. This is th~ true analysis of stems ending in -uu, or -TT. Thus,

, stem Ta/'f, is for Ta/'ff.(J}. TaG'UOJ,

~v>..auuOJ, " ~VAa", " ~vAaic'OJ. TapauuOJ,l " Tapax,. " Tapax'OJ. , ,

" ,

" ,

0pvuuOJ, opVX, 0pvX£OJ.

So JU'sOJIJ for JUS/OJIJ is for original p.f!!'(to,IJ, and 8Uf.TE (stem 0") is for 8"tE. Accordingly uu represents not only Tt, as in many instances, but ali:!o "It, /Ct, X" Sometimes, ai:! in "pa~OJ (stem /Cpa/'() for "pa-ytOJ, "I' passed into ~

§ 3. In Latin, c 2 becomes, several times, g: 88, (1) After n: as in quadringenti and septingenti, compared with du­centi, sexcenti, etc.

(2) Before n. Thus: salignus, willow, from salix (stem, salic), is for salienns, as dignus is also for dicnus; for the proper appreciation of which, compare 8,/C~, U/Ca£or;, and dico (litem, die), and disco.

(3) Before I: as in ncgligo for ncclt·go. (4) Before m: as in segmentum from seeo. (5) Before a vowel: as in negotium for nec-otium. So the Latin lacus, a lake, has become the Italian lago.

In lIuch words as ignarus (= in-gnarus), ignavus, cog­noseo, and ignoseo, the g represents an original guttural be­longing to the simple root in Latin, but now lost: as in

1 The form 'Tf'TP'lXllt fonner!y thollght to have come from an imaginary verb "'""XIII, WII8 shown by Duttmann to be derived from 'Tap4aVOl.

x C and g were pronounced hard in Latin j c like our k, and i as our hard g.

276 Comparative Phonology. [APRIL,

gnosco, the archaic form of nosco am) gnavus, of navus. Compare 'Y'-yvW(T/CfJ) (stem ')'VfJ)) and 'YEwa1o<;.

§ 4. G becomes c before t: aR in Ilctum and rectum, from ago and rego j c (for Ie) being the smooth mute with t, another smoot.h one.

II. LingualR. These are, in Greek, T, S, 1;:)-, A, v, p, CT j and, in Latin, d, t,

th, ), n, r, s. lst. The Substitution of Greek Linguals for each other. § 1. The liquids nre interchangeable with each ot.her: a!1,

(1) A and p. 'rhu!!: lCe<f>o,MP'Yla, headache, is for lCet/»>..a:A­'Y{a, and ciP'Ya).£o<;, difficult, is for aA'YaAeo<;. Compare, si mi­larly, Epe{3w';!to<; and :>..€{3wl;:)-o<;, pulMe j CTV'f,jAO<; and CT1'Y'IpO<;, si­lent. j AeipwlI and Lat. \ilium, a lily j and also Lat. rumpo, perf. rupi (stem, rup), and Sansk. lup and lump, to break. In the same relation stand lp'<;, strife, with Lat. lis; mille and millia, a thousand, wit.h p.vpw, j gramia, a humor in the eyes, with 'Y:>"o'p.7J. So coerulus, from coelum, is for coelu­leus. In French, similarly, r often represents the Latin I: as in epitre from epistola j ap()tre (apostolus) and rossig­nol (lm,ciniola). Gibraltar is said, likewise, to stand for ge­bel al Tarik, the mountains of Tarik. So, our English word frock is derived from a Middle Latiu word flocus, a monk's garment. The Latin peregrinus (per-ager) is the Italian pelegrillo, French pelerin, German pilger, Eng. pilgrim; so that peregrinate and pilgrim come, immediately, from the same root. In the English word purple (Gr. 7rO~vpa, Lat. purpura, Fr. pourpre), we have a similar substitution of I for r.

(2) A and v: as 7rAeVp.fJJV 1 and 7rJIevp.fJJv, the lung8; AiTpoV and lI{Tpoll (Lat. natrum), soda. In double form8 of this sort, the Doric had a preference for the II, and t.he Attic for the :>... So compare Lat. lympha and nympha, water, with vVp.4n1. Anc:ent Panormus, in Sicily, is now called Palermo j and the name of the modern Bologna was, originally, Bononia.

The Spanish nive! and French niveau, correspond, in the

I In the Latin pulmo(o) there is a metathesis of the I.

o

1860.] Comparative Phonology. 277

same way, with the Lat.. libella, a level; as do the Latin lu­tra and Spanish nutria, the otter, and the Latin venenum, poison, and its Italian equivalent veleno.

(3) p. and JI: as JUJI louie, and Doric JlW, in the sense of mOJl. So com~are /lIt} and Latin ne, not; pJiJJI, whether, and Latin num; and also Sansk. accusative suffix -am, Gr. -Oil, Lat. -urn. Final JI, in Greek, is generally an alternate for "., as in the lst pers. sing. of the imperf. act. lTV7I'ToJl for fnnrrop.(,); and in the acc. sing. ending JI of nouns; but sometimes it springs from a, as afterwards shown.

(4) JI and p. Compare t.he masculine comparative forms of Latin and Greek adjectives: as O,ICtO)7J and ocior, fUl~o)JI

(for p.eylO>JI) and major; also &w~ and dims, terrible, and _poll and donum, a gift.

\ 2. The other linguals (the dfmtals and sibilant, which is but the dental aspirate) are interchangeable with each other.

(1) A radical 8 or 'I' before " becomes generally a, and sometimes ~, while in Sanskrit it remained unchanged; as in '1f'AoVa~ for '1f'AoVTWO;- (from '1f'AoVrOo;-) and ovala, being, essence, for 07J'l'la, and 'YEpovata, a senate, for 'YEpovrta. So, EtoJUU (stem, l8-) is for l8lop.tU, with which compare Lat. se­deoj and ~~O) (stem 08-) is for 08/.0, ; and OX'~o) is for OX£8loJ. ]n a few cases, double forms of the same word in -a£a and "I'£a exist, as in Jlavala (from Jlaw, a ship) and Jlo,V'I'la, Lat. nausea, sea-sickness. ..

The change of 'I' to a, in feminine adjective and participial forms, originally ending in -'I'£a, is especially interesting. Thus the fenfinine suffix -aa, of participles ending in -o)JI, -ao;-, and -EW, as TV7I'T6vaa, 'T1n[riiaa and 71Xf>~Eiaa, stands for -'I'£a.

The proper feminine ending is here, as in ";8w (stem ";&-) , that of -£a; and the final letters of the stem are, in each case, -!'T. So that

'l'V'lf'ToVaa is for TV7I'T07J'l'-£a, originally . .",yaaa 1" 'T1n[ra7J'l'-£a, " 'I'Vf/J~Eiaa" 'I'Vf/J~hrr -£a. "

I So in ... ii., ... a.rtl, ... ii .. we have represented with several changos the following original forms: ... ii. for ......... (.) - .. an for ........ 14 - .. iiI' for ... ""T. 'AI'IIUT~lTltl, li~ewiae is for ~itl. from adj. ""/uIT~&' j ."tlITi is for """","i, in which origi-

VOL. X VII. No. 66. 24

278 Comparative Phonology. [APRIL,

The true analysis of the changes that have occurred in the above forms, is the following: T was changed to CT before " and the, afterwards rejected; while also v was, according to uniform Greek custom, cast away before CT, and the previous vowel was lengthened by way of etymological compensa­tion. Stems in -EJIT preceded by a vowel, as xaptE'<; (stem Xaplevr) for x,apI£JIT<;, have, in the fecninine, the ending -ECTCTa. for the original -WT'w.. Here, not only is -T£a changed to -CTa., as above, but v also, instead of being dropped, is assimilated to it, and changed to <;. In such feminine forms as

p.e'Muva, of p.e~ (for p.eMJJ<;) stem p.e>..a.», TaMwa, of TaM<; (for TaXtwr;) stem Ta>..a.», T~PE£JIa, of T~P"1JI (for TEp€ll<;) stem Tf,pEJI,

the same feminine ~uffix, -'4, really exists, but the, is placed, by metathesis, before the final letter v, of the stem, because probably, as that is one of the strongest of all the consonants in itself, the Greek ear forbade its being weakened in the feminine, compared with the other genders, by having two vowels after it, one of them the soft,: so that ~va repre­sents an original p.e'Mivw..

(2) The sibilant (CT) is also interchanged with T, in many forms where it would be final, and in some, also, where it would occur initially. For the exchange of CT for T final, compare, with 7rpo<;, the Homeric form 7rpOTl (Lat. prod-), Sansk. prati. So the neuter suffix -0<;, of the perf. participle acti ve, as in TETIJI/>O<;, is but an euphonic form of the radical • -OT, as the masculine suffix -(0)<; (TET1l4>W<;) is also of -OT<; (TETV-#T<;). SO al~o the final T of those neuter stems which end in T, and do not, like CTilJp.a, drop it in the nominative, is changed, in that case, to <;, as in T~pa<; (stem T~paT) and ICE­pa<; and ICpea<; (stems lCepaT and ICp~aT). For the exhange of· CT for T initial, compare TEO<;, epic and Ionic form of CTO<;, with the same j also, Sansk. tvam, Lat. tu, and Gr. trV lind the

nal form (-,..,.,) of the third pcrs pI. act. of all verbs in Greek (cf. Latin third pers. pI. ending in -nt), the principal tenses are still found ending in the Doric dialect throughout. Compare also Attic dKO.,.'. tll'onty, with Doric FfUcCITl. Latin viginti, Sansk. vin~ati.

1860.] Comparative Phonology. 279

Cretan TP~ (for TFe, Sansk. tvam) with u~, ace. case. Com­pare, also, T&U~, so great, and Lat. tot and totus; and also TEK.o<; ofti&pring, and Lat. secus and sexus.

The interchangeableness of T and u, both phonetically and graphically, is a fact very noticeable in the pronunciation and orthography, one or both, of almost all languages. The in­terchangeable spelling of the Latin adjective suffix -tius, as such, or as -clos (as in adventitiul:! or adventicius), and so of the nom.jnal suffix -tio, as such, or as -cia ( as in condi­tio and cOndicio), is noticeable in this direction. So, in the modern languages generally, t before i, in the same syllable, has a simple or mixed, s-sound. Thus, in Frencb, nation is pronounced as if nl1sion; in German, as if nab-tsi-one; and, in English, as if DB-Sbun.

In the Laconic dialect, even ~ was often changed into tT,

as in uw; for ~eo~, a god, and IuytUTo~ for luya~o~, good, ua'Jvur­ua for ~aA.ao-ua; and 7TauOp and 7Tluop for 7Ta~o~ and 7T"~O~ (when, also, ~ final is changed euphonically to p).

(3) An original sibilant is also, itself, sometimes repre­sented by JJ final. Compare ~JJ, he was, with the Doric ~ and the V~dic as. So, in the 1st pers. pI. pres. act. of the verb, JJ final stands for ~: 8S in rfnrToJUV for T-fnrrO}U~ (Doric form); witb which compare the corresponding suffix -mati, in Sanskrit, as in dadamas, we give, and the corresponding Latin form in -mus, as in dam us, we give. The Greek dual suffix -TOJJ is the equivalent of the Sanakrit -tbas. • A,~, poetic form of alet(Ellg. aye), always, is, in Doric, aUv.

2d. The Substitution of Latin Linguals for each other. § 1. D. (1) D is sometimes substituted for t, especially

before r: as in quadraginta for quatraginta, aud quadratus for quatratus. So, the ancient Mutina is now Modena; the river Atbesis, of old, in Italy, is the present Adige; and Padua represents the ancient Patavium.

(2) Other letters are, in several cases, substituted for an original d: as,

(a) R, in one case: meridies is for medidies (medius+ dies), noon. So r, in parricida for patricida, is equivalent in one case similarly to an original t.

280 Comparative Phonology. [APRIL,

(b) L, also, represents: in some cases, an archaic d: as in lingua, archaic dingua, and lacrima, archaic dacrima (Gr. Mttpvp.a). So, the Spanish cola, a tail, is but another fonn of the Latin cauda.

(c) B, sometimes, represents an original d followed by u or v: as in bonus, archaic duonus j bellum, archaic duel­lum j and bis for dvis (Cf. Gr. 8t" for 8Ft .. ).

§ 2. T. (1) T often becomes s, after r, as in the supines of many verbs. Thus tersum, mersum, cursum, v~rsum, and other supines in -sum stand for tertum, mertum, etc., accord­ing to the analogy of the regular supine formation in -tum, of the various conjugations. The liquids, in fact, generally, ex­cept m, evince a special fondness for having s succeed them.

(2) T is in one case interchangeable with r, as in parricida. § 3. S. S is readily interchanged with rj as in arbor and

arbos, honor and honos. The archaic forms of plurimus and melior were plusimus and melios, as ill meliosem. Corpus (stem corpor) is for corpos, and this for corpor; and genus (stem gener) is for genes, and this for gener. So the I.aco­nians often changed 0" to p in the end of words, as in Ttp for Tt", and 'Trap for 7foV<;. Ancient Massilia has become similarly the modern Marseilles. In German, a like interchange­ableness of rand s is noticeable in the wordt! darum, there­Core, and warum, wherefore, which are compounded oC urn + das, reversed, and um + was j as in English therefore stands for that-Cor j and whereCore for which-for. Compare in the same way, German Hase and English hare j German Eisen aud English iron.

III. Labials. These are in Greek 7f, {J, 1/>, and p.; and ill Latin p, b, v,

f, ph and m. As the changes and substitutions that occur in them

belong, almost all, to the class of assimilative 8ubstitution~, they demand no full, distinct treatment here, except in the following general particulaTl! :

§ 1. In Greek, initial p. is sometimes interchanged with {J, as in fJ'AlTTEW for P.xtTTEW; {AWO"ICCJJ for P.NiJO"ICCJJ; and {Jpo-r~ for p.PO'To<;.

1860.) Comparative Phonolol!'!J. 281

§ 2. In Latin, v becomes u, or is vowelized before a con­sonant, as in cautum for cavtum (caveo); fautum for favtum, and lautum for lavatum. As b and v, like p and 'f or ph, are all correlated labials of but different degrees of bardness, the substitution of u for b in such words as aufero and aufugio, for abfero and abfugio, is of the-same sort.

§ 3. One of the most frequent of all correspondences and interchange:! in different languages is that of gutturals and labials. one with the other. Labials in Greek often corres­pond to gutturals in equivalent Sanskrit and Latin forms, as in ewop.a£ (stem hr. for um.), to follow, compared with Sansk. sach and Latin sequor (pronounced as sekor), root seq.; r'1l"7r~, Aeol. iuo~, and equus (as if ekus); 7revre and quinque (as if kinke). So the interrogative and indefinite words 7rilJ~. 7ro-re. and 7roi'o~. are in the Ionic dialect "ilJ~. "oTe. and ~~, corresponding with the Sanskrit kati, kada, etc.

§ 4. In a few cases, also, Iinguals and labials interchange in different languages, especially T and 7r (t and p); as UTa­&ov. Doric tT7I'aowv. Latin spatium; and so tT7I'eVSQ) and Lat. IItudeo.

I V. The aspirate H. The Latin h is a much harder aspirate than the Sanskrit

b, which it sometimes represents. Before s they both be­come x j as in vexit from veho, Sansk. aviJ.kshit from Yah, to carry (cf. Greek Ox,em). In 11axit, perf. of traho (perhaps for b'a-veho), the same fact appean.

(2) Assimilative Substitution. Assimilation is the result of a strongly determinative, phonetic attraction between one consonant and another, when in immediate juxtaposition. The law of assimilation commonly works backwards, or from the second consonant to the preceding one. as in WvIJJU for ItTJllJJU (for HUJfIJJU, Lat. vestio), and jussi, perf. of jubeo, for jubsi. But sometimes the law works forwards, from the first consonant to the second, as in IJUIJJU for 8")WIJP', r;:rappo~ for~&pcr~. and ~ for J,N'(1i, Sansk. anyas, Lat. alius. So when 'll'pOuQ) was changed by metathesis to 7ropum in the Attic dialect, it was ere long harmonized to 7roppm. Posi­tive full assimilation is the literal change of one consQn~nt

24-

282 Comparative Phonology. [APRIL,

to the same as the other connected with it; as in suffero for sub-fero, and iHatus for in-latus. A more incomplete assimi­lation occurs in the change of one consonant, in juxtaposi­tion with another, to one of the same class with it j as in imberbis for in·berbis, and impertio for in-pertio: m, band p being alliabiais. In nihil for ne-hilum, and nisi for ne-si, and bubus for bobus (for bovibus), we seem to have a few cases also of a retrogressive vowel-assimilation.

I. Gutturals. The law of harmonization is the same with them, as

with all the other mutes, in Greek i that smooth mutes must combine with smooth, middle with middle, and rough with rough i except that, in reference to the rough mutes, there can neither be a duplication of the same mute in juxtaposition, nor a repetition of it even in successive syl­lables. $acf>rp&J is accordingly changeq to $a'TT'rp&J, and BaX­xo<; to B&.it:x.o't, and T~J.U takes the place of ~~J.'£, and 'lrE­

t/>t>..'1J1Ca of t/>Ecf>i">..'1JICa. Before J.' a guttural of whatever degree becomes uniformly 'Y, or medial. Thus OUl)"~'t becomes oUJYyp.O't, and {3eppEXJ.'4' becomes /3e/3pe<y}-'4£.

II. Linguals. 1. Greek. § 1. The Dentals, '1', 8, ~. (1) Before dental mutes, other dentals are changed into

the semi-vowel fT; to which Pott, Curti us, and Heyse agree in gi ving the appropriate name of dis-similation; so that aVV'T'T'o't becomes avvfTTo't; ao.reoJl becomes &.trr6oJl, and 'TT'n~­~Jla£, 'TT'E£fT~Jla,"

(2) Before J.' a dental becomes fT; as in rfTj.'EJI, fil"Rt pers. pI. of o~a for tsJ.'EJI, and i1vllfT}-'a£ for ;Jvvrj.'a£, perf. pass. of aJlVrt». Attic form of aJlUc".

§ 2. The Liquids. (1) L. (a) The weak vowel £ (or y) originally succeed­

ing ">.. in many forms was afterwards converted into A. as in p./i).)..ov for ~MoV, compo of }-'4M; Q)..Ao<; for ~; IJ.">"Mp.tU for a">..to}-'4£ (Lat. salio for saliomi) i fTTb.N" for fTTE">..u.,; /3A­N» for /3a">..(m.

(b) In the AeoUc dialect fT was assimilated to a preceding

1860.] Comparative PhfYo/llogy. 283

~ as it was indeed also to p, II and p. We sometimes find this same style of assimilation in Homer, as in r:,4>€XM for ~Ma, Attic ~'M, first Aor. of o~. In the Attic form the tense-characteristic CT is rejected; and the preceding vowel e is lengthened by way of compensation.

(2) N. (a) Before A. or p, II is changed into the same liquid, as in CTVXACY'f~(JJ for U1JII-}.qyt~(JJ, and EpplJlfiJ for EII-plll(JJ. (b) Before p, II is not thus changed, as in Ellpt'fr'r(JJ, EllpTryJll)P' ; except in words compounded with CTVII, as in CTVpP{(JJ. (c) Before CT, II is, in the word u6v, a8similated to CT, as in CTVCT-1TaJc." or dropped, as in CTVlT'1rcW (CTVII + IT'1rcW). In 7raMII, also, we find II changed to CT in the compound 7raXkrovro<;.

(d) Before a guttural, II is always written "1, as in uvrylCa~m. If the guttural is itself "1, then it is to the nasal gamma (Eng. ng final) that JI is converted, as in CTV"f'Yan7<; (CTW + ~).

(3) S. Before the dentals, and the labial liquid m, any dental may bt! changed into CT, as in olu~a for ol~a, and OO-p.rj for ooprj. So, in Latin est, he eats, third pers. Sing. of edo, for edt (for edit), we have d turned to s, before t.

2. Latin. The Dentale. (1) The dentals, d and t and the liquid r, are sometimes

before s assimilated to it; as in cessi, perf. of cedo, for cedsi, gessi, perf. of gero, for gersi, concussi for concutsi, possum for potsum, Meum for fidsum, for fidtum, and missum for mitsum, for mittum. Such perfects as sedi, fidi and sCIdi, with supines in -ssum are undoubtedly contracted forms of original perfects in -si, as sedsi, fidsi, etc.; from which after­wards the s was rejected for better euphonic effect, and the short radical vowel, e or i, was lengthened by way of com­pensation.

(2) D was sometimes assimilated to I before I: as in sel­la for sedla, for sedela from sedeo, to sit, and lapillus (for lapidlus) for lapidulus.

(3) N was assimilated to I, m, and r: as in illino (in+ lino, immineo (in+mineo), irrno (in+ruo).

In some of the modern languages, especially the Italian,

284 ComPf!rative PkoTlOlngy. [ApRil.,

the law of assimilation is quite active: as in Ital. atto, an act (Lat. actum) ; patto, a pact (Lat. pactum), fitto,l trans­fixed (Lat. fixus).

The letters most frequently doubled by assimilation, in the middle of words, are the liquids.

III. Labials. lst. In Greek. ~ 1. M. Whenever a labial precedes II-! in the middle of

a word, it is changed to p-: as in "Ipap-p-, for "Ipatpp.t] from

"I~' • § 2. n, B, ~. These all, when preceding tT, combine with

it, into the compound consonant 'F; which, while having, analytically, either one of the labials for its base, has yet, tQ the ear, always the sound of the smooth mute '1T'. So, in Latin, scribsi, perf. of scribo, becomes scripsi.

2d. In Latin. M is, in a few cases, changed to n: as in tunc for tum-ce;

princeps for primum (ee. gradum) capio: clandestinus, adj. formed from clam (for celam) ; tandem (from tam); and so quanquam, eundem, etc.

The interchanges of the different labials, one with the other, in various languages, may be here advantageously re­called: as in SANSKRI'r. GREEK. LATIN. GERMAN. ENGLISH.

upari. Vrrep. super. liber. {over.

:~:::.. hyper.

saptan. ma.. septem. sieben. seven. 2dly. Topical Substitution. By this is meant a change of place, in a letter or syllable,

either by accident, if there be any accidents in language, or for better euphonic effect. Topical substitution is of two kinds:

(1) Metathesis. (2) Hyperlhesis. (1) Metathesis (from /UTaTCJ'I]P-£' I exchange) is a change

in the order of the letters of a word, in the same syllable. lst. In Greek.

1 Webster's reference to peto, as the etymological radical of the noun fit, .. absurd.

1860.] Comparative IJJ&onology. 28fj

§ 1. It occurs in several, separate, individual words, that bave no common elements of classification, unless it be that the consonant, before and after which the vowel plays inter­cbangeably, is a liquid (p): as "aproe; and "paToe;, strength; ~tlpa-~ and ~p4To<;, courage; ~pdxrlCo> (stem, ~op),I leapi ICpa-8£", and 1CILp8la, the heart. In Homer we find both lCapTEpoe; and "paTE~, strong. np6uo> became, afterwards, 7T'Opuo>, and, still later, 7T'dppo" Lat. porro. So, compare Aeol. TipToe; (Lat. tertius), Eng. tierce and tier, with Tp{TO<;.

§ 2. It occurs, frequently, in the perfect of verbs whose stems end ill a liquid: as TETiJI'I"a from Tip.JIO> (root, TEP. or TI.If'), (:JifJA",,,a from fja;A."MJ (stem,fjaX). TE~"a from ~V7]UICQ) (root, ~",,). Compare, also, the perfects of ICQ,}.Jo>, JeJ.p.JIO>, etc.

2d. In Latin. § 1. A few cases occur, in proper Latin forms, compared

one with the other: as, tero, perf. trivij sterno, perf. stravij ferveo, supine fretum, cerno, and cretum, sperno and spre­tum.

§ 2. There are, also, a few cases of metathesis,l in equiv­alent forms to certain Greek words; as UICf7T'TOp.a" I look around, and Lat. specio; "pwo>, I judge, and cerno; ¥o> and spuo, I spit. .

(2) Hyperthesis. This (derived from inrepr/:;!r",,,,,, I place or carry over) con­

sists in changing letters from one syllable to another. lst. In Greek. ' § 1. This occurs in a few single words: as in the geni­

tive of nwE, the Pnyx, nVICrlOe;, which case, from its resem­blance to the adj. 7T'V1CrIOe;, crowded, shows us the undoubted etymology of the word. Compare f¥x~ for &A.xoe;, the peo­ple, Cretan 7T'oXXO<;, Lat. vulgus, Germ. yolk, Eng. folk.

§ 2. Many verbs, having now the diphthong E' in their stems, exhibit therein a change of place of the weak vowel..,

1 In English, an orthoepical metathesis often occnrs, if not an orthographica!; •• in the prononciation of iron and fire, and in the ntterance of the aspirate first in its combinations with an initial w, &I in soch words &I which, what, where, ICC.

286 Comparative Phonology. [ APRIl.,

which originally followed, instead of preceding, the final con­sonant of the stem. Thus:

TfLvr.J), stem TW, is for TWu". ')(,ftpwv, " ')(,fp " ')(,fPu",,· ap.elvwv, "

, " aJU"twv. ajUII,

{:Jatvw, " {:Jav, " {:Javle" (ef. Lat. venio). p.alvop.at, " p.av, " p.avu"p.tU. </>atvw, " </>av " </>avlw.

§ 3. Several feminine adjective forms in -cuva exhibit the same change: as ,d"NuJJa for JUNW.£a, T~~ for Ta>..tivl4l, etc.

2d. In Latin. There is, in the word nervus, in Latin, as the equivalent

of vroPOII, a single instance of hyperthesis, in the one lan­guage as compared with the other.

So Bosra, in Africa, now represents the origina.l BVp<T40 Some French derivatives from the Latin, exhibiting the fact of hyper thesis, will not be inappropriate: as, tremper, to temper, Lat. temperare; tout, all, Lat. totua; noeud, a knot, Lat. nodus; peuple, the people, La.t. populus. Iu raison (ratio), maison (mansio), palais (palatium), we ha.ve un­doubted instances of the same sort, in which the " is to be regarded as radical, and not inserted, as in faim (fames) and foin (fenum), as a diphthoogal compensation for a shorten­ing of the original form.

In the case of some aspirated forms, there occurs a curi­ous transfer, not indeed of a letter or syllable itself, but of a special affection belonging to it: as in ~~ fut. of T~ lEw fut. of lxw; 7rtUrxw (for 7ra~Q'~w), stem 7ra~; and ~pIE, gen. TP"'X0'>. Compare, also, ~fQ'p.O,> with Doric Tt!i)p.O,>, a stat­ute. Here the aspirate, when lost in ODe part of the word by contraction or flexion, is carefully borne, for preservation, to another part.

The next style of Consonantal Changes consists : B. Of Insertions and Additions. These are of a threefold

character: 1st. Prosthesis. 2d. Epenthesis. 3d. Epithesis.

o

1860.] Comparative P/&01IOiogy. 287

1st. Prosthesis. This consists in prefixing a single letter or 8yllable to the beginning of a word, and for the purpose simply, in nearly every case, of better euphonic effect.

§ 1. The vowel prefixes of a prosthetic sort, in Greek, are a, e, and 0, and, once or twice, £.

(1) e&. Compare a,u~, to pluck, with its other form ,u(FtOJ, and so a,up&" and ,up&1, to bereave; a~ and Lat. mulgeo.

(2) e. Compare lpv~pd<;, red; Sansk. rohita, Lat. raber ; Ex~E<;(also ~~), Sansk. hyas, Lat.·heri for hesi. In ~~and ~~ we have both a fuller and contracted form of the same original word, in which the e is radical and not prosthetic.

(3) o. Compare 8oov; (for &Sovn) , Sanek. dantas, Latin dens for dents; 8J1Op4, a name, Sansk. naman, Lat. nomen; OJU')(,EOJ, Sansk. mih, Lat. mingo, I void 'vater.

(4)" As laVa,; I sleep, compared with aiNs. § 2. The letter a- is found initial in lIome words, which

appear at other times without it: as in a-p.ucp6<; and p.I./tp/w, 'T'/:tyo<; (Lat. tectum) and tTTe.yo.;, a-p:upawa and p.vpawe&. In some cases where a- thus occurs, it is radical to the original form; and in some cases it may be, possibly, the frag­mentary representative of a lost prepoeition (el<; or l<;), serv­ing to give the form to which it was prefixed a more strongly directive sense; just as, in words beginning with ."".., 11-, and a-, we often have fragments of an otherwise lost privative dvEtJ. Other prosthetic additions, particularly e, may have sometimes originated in this way, and be but the remains, occasionally at least, of a primitive prepositional prefix.

Prosthetic additions to the original radical elements of a word often occur ill French and Spanish. In Spanish, as in French, e is prefixed to words derived from the Latin begin­ning with sc, sp, and st: as in

SPANISH.

escribir, espeso, estado, 1

PRENCH.

ecrire (originally, escrire), epais, .itat

LATIN.

scribere. 'Jpissus. stat.us.

1 In this way English orthography has been complicated with French-Latin fom.s of Latin words, 81 in estate (Lat. status), etpoatle (spons." especial (spe­cies), establi.h (Itabilio).

288 Comparative Plumology.

2<1. Epenthesis. This is the insertion of a letter in the middle of a word, for

the purpose of a better dynamical or musical effect. In the Sanskrit, llfter the prepositions sam, ava, pari, and prati and some words beginning with k, an euphonic s is intro­duced between them and the words with which they are compounded. With this euphonic use of s, a similar addi­tion of it to ab and ob, in Latin, before c, q, and p, re­markably agrees. Ob sometimes retains it even when alone.

lst. In Greek. § 1. ~ has an affinity for T, ~, and p., and often occurs

before them, after short vowels: as, in the 2<1 pers. dual and plural person-endings passive of verbs, before ~: as, .¢.)-o ...

and -o"~E ; and in the 3d dual passive person-ending of the historical tenses -u'ih,I' ; with which compare the correspond­ing person-endings TOl', TE, and T'I'II', in the active voice.

§ 2. We find also, in Greek, other epenthetic uses of dif­ferent consonants: as, (1) Of fJ after J.' ; 8:s in J.'EO""Ip.{Jpla (= plO"'TJ-t~plpa) and ~aJ43~, astonishment, compared with ~avp.a, wonder. In French, a similar fact appears in some words: as in chambre, Lat. camera; nombre, Lat. nu-merns.

(2) Of 8 after 1': as in a"E~, contracted a",spO~ ; with which also compare Fr. gendre and Lat. gener.

(3) Of ~ after 0": as in Ip,&,u~},."1, a thong; with which compare Ip,&,uO"ro and ~'E. The ~ serves, in such cases, to facilitate, phonetically, the union of J.' or " and 0" with the succeeding },. or p.

2<1. In Latin. § 1. N. ,In the Latin equivalents of some Greek and

Sanskrit words an epenthetic n, or an n inserted for mere euphony, occurs: as in anguis, a snake, Gr.Ex'~, Sansk. ahis. The nasalization of various verb.steml'l, in the prescnt and imperfect tenses of the different voices of the verb in both Greek and Latin, as in fundo, perf. fudi, and KUlIEro, fut. tcVO"ro, will be considered, by itself, under another head; and is therefore not embraced in this lIection.

1860.] OompMative Phonology. 289

§ 2. Pis epenthetically inserted between m and t or s: as in ilumpsi and promptus from sumo and promo. Com­pare Fr. dompter, to subdue, and Lat. domitare j and also t.he English word tempt, and it" Latin original, tentare.

§ 3. R is euphonically inserted, by epenthesis, in the genitive plural, between the stem-vowels a and 0, of the 1st or A-declensjon and of the 2d or O-declension and the proper plural genitive case-suffix -urn: -arum being for -aum, and -orum for -oum j with which compare -6)11, gen. pI. suffix in Greek: as in }WIJO"a-rovrcontracted }WIJO"ow. The r epenthetic, in Latin, prevents the unpleasant hiatus other­wise made by the concurrence of 8+0 in the olle cal:le, and by 0+0 in the other.

§ 4. S is used epen'thetically, with ab and ob, in com­ponnd forms: as in abstineo, abstraho, obstinatus, and obsto. In subscus (sub-t-cudo) compared with incus, we see a similar use of it with sub.

Caution: D, it is often said, is also epenthetically in­serted between two vowell:!: as in prodeo (pro-teo), and in the 2d pers. sing. and pl. of prosum (prodes and prodestis), and elsewhere in that verb. The same fact is cited, also, in reference to red eo (re-f-eo), reddo (re+do), and redarguo. The d, however, in these forms, is not epenthetic, but radical. The Sant!krit original of both forms is prati. Its Greek equivalent, 'If'PO<;, was accordingly, at first, 'If'POTt, in which form we find it in Homer, and from which, T being inter­changed for 0", it became 'If'PO<; by contraction. Prod- and red- are, therefore, nearer their originals than pro- and re·, their shorter forms. In such forms as praeeo and deerro, no difficulty was felt by the Latins, on account of the hiatns caused, as there t!hould have been, on the suppo~ition that d, in the prefix~s prod- and red- it! of a mere euphonic origin.

3d. Epithesis. . This consists in adding a letter or syllable, at the end of a

word, for better euphonic effect. The II 'cf>eX"U/TT£IroIl, in Greek, is an addition of this sort,

which, from its inherent phonetic strength, furnishes a good VOL. XVIL No. 66. 26

290 Comparative Phonology. [APRIL,

staff on which the voice may rest, at the end of a clause or sentence.

No epithetic addition of letters, in t.he modern languages, occurs to the author. There are, however, in French usagE', frequent instances of phonetic, if not of graphic epithesis, in the utterance of the final letters of words which, by them­selves, are silent whenever they are in regimen with words immediately following them, which begin with a vowel. So, too, the cardinal numerali:l, in French, which end with a con-

080nallt, as six, sept, huit, dix, have their last letter, other-wise silent except before a vowel, distinctly pronounced when at the end of a clause or sentence.

The third class of Consonantal Changes we term: o C. Suppressions and Abridgments. These may occur in

the three different parts of a word: its beginning, middle, nr end. Such suppressions are denominated, according to thpir nature and position, by the following different namE'~: aphae­resis, elh;ion, syncope, ecthlipsis, and apocope.

lst. A suppression of a letter in the begillning of a word. This is termed aphaeresis.

I. In Greek. § 1. ~ often vanishes entirely, in Greek, at the com­

mencement of a word; or, more frequent.ly, is replaced rather by an aspirate, when a vowel follows. Sometimes both forms occur, as in aV~ and ~ (Sansk. sukara (s), Lat. sus, Germ. sau and schwein, Eng. sow and swine.) So also UUAauUG

(Doric) and ~aAaaua, the sea (for llAaaO"a) from liA~, salt, Lat. sal, Sansk. sara, salt. In respect to ~aAaaua and a:\.~,

compare /Lp.a and ~ap.a. The Romans liked the IE'tter U

much better than the Greeks; and the aspirate is, accord. ingly, often initial in Greek where, in the equivalent forms of the Latin and the Sanskrit, the sibilant occupies its place: as in ma, seven, Lat. septem, Sansk. saptan; and EE, Lat. sex, Sansk. shash.

§ 2. In a few words :\. was dropped when initial: as in Uc~ for :\.u,~, to winnow, and ty~, mortar, for >.1~. So, in the Aeolic dialect, p. was dropped from p.fa, one, which thus became ia.

o

1860.] Comparative Plwnology. 291

II. In Latin. § 1. We often find s suppreflsed or wanting, initially, be­

fore other consonants; which is retained in the equivalent Greek forms, as found in the Sanskrit; or, in some cases, prosthetically applied in Greek: as,

LATIN.

!ego, I cover. fallo, I deceive. fides, a cat-gut. cavo, I hollow. vespa, a wasp.

a-reyo> U~

GREEK.

urp~ (cf. rpel&p4£) UICO.'1f'To>

ucf»1E

SANSKRIT.

sthag. sphal. bhid, a filament.

§ 2. There are Borne interesting cases of aphaeresis, in in­dividual Latin words:

LATIN. GREXK. SANSKRIT.

80m is for esumi. Cf. elpJ (for Eu,.,.,) asmi. nosco" gnosco. "'Y£'Y"o,UICO> jni. (desiderative form, navus" gnavus." 'Y£1IJIa'io~ [jijni.sft..)

So the English word stranger (Lat. extraneus, Span. estran­gero, Fr. etranger) has lost its initial e: as in estrange; as alBO the word story (Gr. ia-ropta, Lat. historia, Ital. istoria and storia), has lost the initial syllable hi.

2d. A Imppression of a letter or syllable, in the middle of a word. This is called by different names, according to cir. cumstances.

§ 1. Elision. This occurs when a v:owe1 is removed from beCore another vowel: as in nullus (= ne+ullus), nunquam (= ne+unquam), aveP'Xop4£ (= tlll&+E'P'X0p.a£).

§ 2. Syncope. By this is meant the removal of a vowel from between two consonants: as in patris, gen. of pater, for pateris; and so 'TT'Q:rpOr;, gen. of 'TT'anjp, and Homeric TL7rTe for Tl'TT'OTe. Valde, in Latin, is for valide, by syncope.

§ 3. Ecthlipsis. This is the removal of a conllonant, or oC an entire syllable, from the middle of a word.

1st. In Greek. In Greek, U is often rejected by ecthlipsis ; sometimes in

Comparative PluYnolof!'!l. [APRIL,

nouns, and sometime!; in verbs: as in 'Yw£or; for 'YEVEtTOr;, gen. of 'Y~vor;, and j3ov~ for j30tJMVE{u for j30tJMVEtT(1.f,.

While in Sanskrit euphonic principles ruled with a force greater than in any of the cognate languages, still many harsh combinations were allowable, which seemed to the Greeks and Romans, even when occurring in a regular way, altogether too dissonant. In the case accordingly of verbs, having rook! terminating in a consonant, it was an almost universal rule, in both Greek and Latin, although not in Sanskrit, to connect the personal terminations with the stem, by means of an union-vowel. In the following roots, how­ever, the connecting vowel was suppressed, when the personal ending was affixed: in Greek, the roots er;, to be, and w, to know; and in Latin, es, to be; fer, to bear; vel, to wish; and ed, to eat; so that we. have the forms ecrrl, EtTP.EV, LGTE

and f8p.Ev, and also est, he is, fert, vult, and est, he eats. As in Sanskrit, before the personal terminations beginning

with t, th and dh, roots that end with a consonant other than n reject 5, in order to avoid a harsh combination of three consonants; 80, in Greek, roots terminaf.ing with a consonant abbreviate in the perfect passive the terminations -tT~OIl, -~e, to -~Oll and -~E, as T&tJt/>~e for T&,*~e, and Tba~E for T&a~e. Compare in Sanskrit the form stha, to stand, with itself as it is when compoullded with the prepo­sition "ut," up, as in utthita, upstood for ut-sthita.

Before tT the dentals and the dental liquid II are dropped ; as in Mf''1T'ar; for Mf''1T'a8r;, "Opll'; for "oPtJ~r;, tTwJl4O'£ for tTM­p.aTtT" and 8olf'OtT£ for &dfWVtT£. In '1T'OW, stem '1T'oo, not only is tT dropped, but 0 is lengthened also by way of compensa­tion, ail likewise in the perf. act. participle in -~, as in j3ej3olJ­MvttWr; for j3ej3otJMvKOTr;.

When both a dental and II are omitted before tT, the absorp­tion is indicated by an elongation of the vowel, if a, or by it."! diphthongation, if e or 0; e becoming in Fluch a case e" and 0 becoming ou and fI); as in '1T'{UT£ for '1T'aVTtT£, and tT'TT'eUr6J, fut. of tT'TT'€v&', for tT'TT'EvMfI), -rvt/>~e(r; for 'T'tJt/>~evrv, }.£fI)'IJ for AEoVTr;. and 00011'; for l$&vrr;.

2. In Latin.

"')0 [.

1860.J Comparative Phonology. 293

Abridgments by ecthlipsis, accompanied often by a subse­quent contraction, are numerous: as, debeo and praebeo for dehibeo and praehibeoj promo and sumo for pro-emo and sub-emoj malle for mavelle (= magis and velle); prudens for providens ; amavi and docui for ama-fui and doce-fui; lumen for lucimen; hodie for hoc die j judex for jus-dex, and momentum for movimentum. So the dative and ablative pl. suffix-ending -is, is a contraction, in the different declen­sions, of the original forms -abus, -obus and -ibus j with which compare the double dative pl. forms, queis or qui bus of the relative pronoun qui. So poematis is found in some authors for poematibus.

'I'he above instances are of an individual sort, and better denoted by themselves, than by any attempted classification. The facts which remain, that are worthy of note, may be thus classified:

(1) D is often suppressed before 8, and so sometimes is t; as in divlsi for dividsi, mlsi for mitsi, clausi for claudsi, and laesi for laedsi. In divlsi and mlsi, or any such case, the first vowel i is long by way of contraction, as it would other­wise be made by way of compensation.

(2) C, g and q sometimes disappear in the same way be­fore s; as in sparsi for spargsi, mulsi for mulgsi, .and torsi for torqsi.

Even in English, words are sometimes softened by the rejection of a letter belonging to the original root j as in our words speak, spake, and spoken, from the German sprechen, sprach, gesprochen.

§ 3. A suppression at the end of a word is called Apocope. In the Sanskrit, in the final form in which it has reached

us, two consonants were no longer tolerated, as they once had been, at the end of a word; but the latter was rejected. That this feature of the language was not fixed upon it, until after the separation of the other languages from the common parent-stock, would seem evident from the fact, that it is not true of the Zend or of the European languages, old or new. The result to the Sanskrit is a mutilation in the present aspect of many of its original forms, which, if found

25·

Comparalive Phonology. [APRIL,

now as they were in their primeval state, would furnish much valuable light on many etymological questions and theories.

1. In Greek. (1) All final mutes are apocopated from forms, where

they would otherwise appear as a radical part of the word. Thu::I p,EN/r becomes p,EAI.; UWp,aT, uwp,a; Ennn-ET, third pers. sing. imperf. act. (for fuller form ET1J7rTET£) becomes fnnn-€ i and ETIJ7I"'TOVT(£), third pers. pI. of same tense, becomes Ennr­TOV; 'YaMICT' becomes 'YaM, and ~uavr(£) (compare erant for esan!) becomt~s ;juav, and 'TT'o'vr (neut. of 'TT'0,~) becomes 'TT'w,. In such nominatives neuter, as IC€PM, ICP€M, 'T€pa~, the final T of the stem is merely changed to ~.

(2) No consonant can properly end a word in Greek, except v, P or~. Ou" or oUx, is but a mutilation of oua aDd EK or f~ of fIC£~ (cf. EICM, 'ICe'; and eICE'vo~); and they are properly but proclitics, never occurring at the end of a sen­tence; OUIC always preceding a word beginning wit.h a vowel, and fie one commencing with a consonant. In respect also to the three letters, v, p and u, it is to be remembered that p occurs rarely, and that v often represents u, or contains it by ab::lorption, and also that ~ final cannot be preceded by a dental or the liquid v. Even v and u were themselves so weak at the end of words, as to be often omitted. Thus tryw is for f.,yOJJl, Sansk. aham; and TOVrO, neuter of oVr~, is for TOVrOV. Compare also 'TT'pOU"JE and 'TT'pOU~f!JI, vV and , JIVJI.

(3) 'rhe passive person-endings -vral. and -J/TO cannot occur after a consonantal stem ; the v accordingly is changed to a, and the forms become TETU!/>aTaI. and ETerV!/>aTO instead of TETu!/>vra£ and ETETV!/>vrO; like the change of the accusative caflc-:;ign v in the third declension to a, after consonant:!, as in 'TT'aT€pa for 'TT'aT€pV, and ICopv"Ja for ICOpV"JV.

(4) When occurring betwecn two short vowels in the end­syllable of a word, u and v tend to vanish, as in JU~ov<; for JUt'O(V)E~ and 'Y€VOlJf; for 'Y€VE(U)~.

(5) In the third pers. sing. active of all the tenses, and io the third pers. pI. of both the imperfect and aorist tcnses, no original 'T, still preserved in Latin, is dropped. Thus TV7rorE&.

o

1860.] Comparative Phonology. 200

lnnrrE, Trtv4>E, frvtE are for MTET('), £TV7rrET(£), TETJXpET(')' htftET£. Thus compare

SANSKRIT.

bharati, he bears, abharat(i), he was bearing,

GREEK.

~f>E(T)', Ecpef>E( T£),

LATIN.

fer(i)t(i). fere-bat(i).

(6) Un:!igmatized masculine and feminine nominatives, or those which would normally have the gender-sign u affixed, but which, on account of all JI final in the stem, ha ve rejected it, have their stem-vowel lengthened by way of compensa­tion, as in '1TO'J.I~JI for '1TO,~, and .qyepbJJI for .qyEp.OJJ<;.

2. In Latin. The number of final consonants in Latin, c, I, n, r, s, t, is

somewhat greater than in Greek. Apocope occurs in Latin in several interesting classes of

cases. (1) In the loss of the final letters of many con8onantal

stems of nouns in the nominative; as in cor, the heart, for cord; lac, milk, stem lact; Oll, a bone, stem oss (Gr. (loom Oo-rE); leo, a lion, stem leon; and mel, honey, stem mell.

(2) In the ablative singular form of all the declensions; in the dropping of its final characteristic d from them all; as domino for archaic dominod, and :!ermone for sermoned.

(3) In several imperatives, all dic for dice; due for duce j fac for face; and fer for fere.

The next class of consonantal changes is composed of IV; Weakened consonantal forms; or the weakening of

individual consonants in certain specific forms or classes of forms.

(1) The very common one of T into u. Thus the ending -ovu£ in the third pers. pl. of the pres. and fut. active of Greek verbs, as in TVrrTovu£ and T"t0VU£, represents an origi­nal -oJln; which was the form also actually used by the

.Dorians. The analysis of the changes made in the form is this: T was euphonically changed to u, after which JJ was dropped, according to universal Greek usage before u, and the vowel 0 was lengthened, by way of etymological com­pensation, into ov.

296 Comparative Phonology. [APRIL,

So ill Latin, the proper supine-ending -tum is changed, when the stem of the verb ends in a dental, into -sum. After a long medial vowel the dental is thrown away, as in caesum for caedtum, from caedo, to kill, and laesnm for laed­tum, from laedo j as likewise in the supine and participial forms of cado and edo, to eat; in which the vowels a and e are accordingly lengthened by the contraction of the syllable to which they belong, as in ciisum for cl1dtnm, supine of elido, and the participles ambesus and comesus of amb~o and comedo. After a short vowel, the dental is also assimilat.ed to the changed suffix, as in fissum for fidtum, and fossurn for fodtum; supines of fin do and fodio.

(2) That of the conversion of an original u, in the begin­ning of a word, illto the aspirate; as in 11<; for oW, which two forms are both found in use together; and of 'ltrrr,,.,,, for utu­T'rJfJ-' (Lat. sisto). This subject will, however, receive ita proper treatment., under the subsequent head of Sibilation.

(3) That of the weakening of an original Digamma into various forms j another topic reserved for fuller discussion, by and by, alone by itself.

A special hint. It must not be forgotten, that some dif­ferences ill the flexion-forms both of nouns and verbs are to be resolved, not by any mere phonological analysis, but on t.he theory of a manifest duplication of the stems of its dif­ferent forms, and sometimes even by the aggregation of very different stems . toget.her, for grammatical cOllvenience, into one form of conjugat.ion. In such forms as fJ-rya<;, fU!Yt!iII."I. fJ-~'Ya; 7T'o).v" 7T'oU~, 7T'0)."; 7T'pO,or;, 7T'fXU'ia, 7T'ptloll . we have two different original flexion-stems: p.€rya and iU"(&:II.o j 7T'0).t/ and 7T'oX~ j 7T'pQ.o and 7T'pae. The two stem-forms of 7T'0).v, and 7T'0).).~ we filld used interchangeably ill Homer in most of the cases. In the Lat. fero, perf. tuli, supine latum, we have two ab:!olu~ly different stems aggregated, fer and tul; two, not three, as latum is for tlatum (cf. Gr. TMU., and T).'7T~, and Lat. tolero); and datum is from the same root with tuli.

V. Strengthened consonantal forms. Neither learner nor teacher, it is believed, can be harmed

by occasional repetitions of the same fact, in other relationa

1860.J Comparative Phonology. 297

and for other uses. It is difficult, if not impossible, to sur­vey phonology thoroughly on its different sides, and to do jus­tice to each one of them by itself, without at the same time catching views of other parts already examined, or demand­ing afterwards more distinct and complete consideration.

The use of strengthened fonns was one of the early fea­tures of language, abounding in Sanskrit and Greek, and of frequent occurrence also in Latin; but occurring less and less in subsequent and derived languages, as we go in them further and further from their primeval source. As the Latin preserves in most of its aspects more of the simple strong characteristics of the San8krit, than the Greek, its departu!e in this respect tQ a wider degree from its original than the Greek, is to be accounted for probably by the strong practical tendency of the Roman mind, which did not relish double forms of the same thing, and multiplied modes of reaching the same end.

The modes of strengthening stems are various, as : § 1. By nasalization, as in ICap.JIO), stem /Cap., and rtp.7IfI),

stem Tap.; and in Latin, frango, flndo, vinca, compared with their sim pie bases frag, fid and vic. But the subject of nasalization must be treated afterwards by itself.

§ 2. By the reduplication of the radical syllable or sound. A repetition or reduplication of words and syllables is the

most natural and effective style of emphasizing their impor­tance.1 This occurs abundantly in Sanskrit and in Greek, but much less in Latin. See subsequent treatment of Re­duplication by itself.

§ 3. By changing stems originally ending in one of the" mutes or T mutes, followed by the semivowel, (as "1', It"~ X', T" ~,) into (1'(1' or TT; and stems ending in the liquid X fol­lowed by" as A£ into U. 'rhus: TQq(1'ft), Aa/(1'(1'0), t/>pUr(1'O). AUr(1'op.a' are for the earlier forms TO/'fu." XevICk", t/>p"c{o)' X'Tto­JI4'; as, also, {3aAAo>, p;.XAo>, and O"TeAAo> are for {3aX/ro, p.e]..u."

I This idea lies at tho foundation of some of our most exprCIISivo words, 81

respect, regard, remark; where the idea, as in the word respect, is, that the per­IOn or thing respected is worthy of being 100Jr.ed at a second time, or, again and apilL

298 Comparative Phonology. [APRIL,

and uTe)..lw j with which compare p1l"x'}.,oll, comparative of p.O;NJ, for pO,A.tOll.

The gemination of the final consonant of the stem, in this way, answers a double purpose: that of strengthening the stem, and that also of symbolizing, in a form 1:;0 unusual in Greek, except as a special etymological contrivance, the pre­vious changes that had occurred in these forms.

In Latin, the verbs in -io, of the third conjugation, repre­sent the same class of verb-stems that, in Greek, were changed so as to present a duplication of the final letter of the stem : as capio, cupio, and fugio. In Greek, verbs and adjectives having II or p for the final Jetter of the stem, fol­lowed by the half-vowel " a metathesis of· the £ oc(m~, in­stead of a gemination of the final consonant j as in q:,a{lIOJ, "tern q:,all, for q:,alllOJ, {JalJlb), stem {Jail, for {Jalltw (cf. Lat. ve­vio), etc. The half-vowel i, or y, of the original forms of these various words represents the Sanskrit yft, properly meaning to go, occurring in verbs of what is called thE.' fourth class, in that language, and characteristic, very extensively, of intransitive and passive verbs. But y was not a sound to be found in Greek j and therefore it must either be vowel­ized, as if " or expressed by Borne other assimilated sound. In the adjective termination -w~ in Greek, as in Wyw~, Sansk. yajyas, we see the equivalent of the Sanskrit adjective suffix yas. The Doric future suffix-form -UUAJ, answers, in the same way, to the Sansk. syami. The analogue, accordingly, in both Greek and Latin, of the fourth Sansk. conjugation-ending yami, is £OJ or io (for -taJ.U). From such an original regular form in UAJ, come not only the altered forms -uuw, TTW, -t'Q), '7Vw>, but also those in -atllOJ (for -aJltOJ) and -atpOJ (for aptOJ). Ro, in Latin, the adjective and nominal suffixes -ius, -ia, -iee, answer to the Sansk. yas and ya, like the verbal ending -io to the Sansk. -yftmi.

§ 4. By the epenthetic insertion of u in the midst of the stem: as in ~luryw (Lat. misceo) compared with ~byvvJU; tax,w and luXJl€o~, compared with ExOJ j also E~'OJ with eBo­~", to eat, Sansk. ad, Lat. ed.

§ 5. By adding to consonantal stems ending in 'It' and"

1860.] Comparative Phonology. 299

the letter T, and to vowel-stems ~; as in TlnrUJJ (stem TV'7l') j

Kfnrr6J (1C0'1/") ; ICplnrT6J (/Cpvf3) j 7T'EICT6J (7T'f/c); and TtlCT6J (TflC) ;

and for vowel-stems 1C1ITJ~6J, 1CA,~~6J, 7T'P'T}~6J and U~~6J (stems JCJIf1" /CAe, 7T'ptI, and Uo,). Such forms, in Latin, a~ necto, plecto, Recto, are of the same analytic origin.

Even in the forms of nouns, in Greek, the strengthening of the stem by the insertion of T appears: as in the epic forms 'lfTOM:p.o<; and 7T'TOA,£<; for 7T'OA,fp.o<; and 7T'OA.£<;. Compare, also, 'lfTf:PlJo, and 7T'€PlJo" the heel.

We come, now, to some of the special pathological affec-tions of the classical languages.

A. The Greek .. 1st. Its dialects. Every language, covering an area of any considerable ex­

tent, for a long period of time, tends to break into separate dialects; determined, objectively, by different physical and local causes j and also Imbjectively, by difference of employ­ment, development, and culture. The effects of time and space are as marked on men as on nature, and on the world of language and of letters, as on any part of the vegetable kingdom. In Greece, especially; dialectic developments were of the fullest and finest growth. Had ever a nation, in respect to all physical inRuencf's, so favorable a position, as Buch, for growth in all the elements of inward great.ness 7 She was nursed in the mouutains, among the giants. The air that sh« breathed was full of the seeds of life. In the broad blue sky above, and the bright blue sea below, she saw divine aspects of energy and beauty constantly mirrored to her view. Her eye and her heart were ever invited, by sur­rounding objects, to a perpetual festi val. She laid the beams of her greatness on two continents: combining the stern strength of the one with the soft luxuriance of the other. She sat, as queen, on many waters, and girt around, as with a mantle of stars, with clusters of islands shining about her on every side. .

On no spot upon earth can one be born, to this day, where Nature will bend down more lovingly and impressively over bim, to breathe her life and beauty into all the opening ele-

300 Comparative Phonology. [APRIL,

ments of his being. The mountains and the sea have ever been the two greatest natural teachers of mankind. No pt"o­pie could come into more immediat.e contact with Nature, in either of these forms of her presentation; none ever did so meet her constantly, in them both combined. Rome was, in some respects, similarly accoutred for greatness with Greece: with the mountains behind and the sea before; but it was with no such fulness of preparation; her home was, after all, upon a plain. Greece was, everywhere, a land full of broken and rugged surfaces, of bold shores, of Mhort, dark, rapid, foaming streams, and of every variety of landscape, skirted, at ten thou8and points of contact, with the sea, which not only surrounded it, bufcrept in, with its pulses of ever-quickeningforce,'into all the folds of its physical and na­tional life. Amid such influences, as each vegetable and ani­mal have not only their special geographical zonE', but also even a specific climate and locality, where they will best fill out the whole ideal outline of their being, man finds his mo~t favored spot for a large growth of life and action.

When t.he western fracture was made from the common Graeco-Italic stock, which, under the long action of many fa­voring circumstances,'was perfected, in the end,into the round orb of Roman life and law, the portion remaining behind, within the boundaries of Northern Greece and Asia Minor, began slowly to form a local character and language, as they settled more and more upon the same soil, into fixed COID­

munities and habits. No language can bear greater evidence of home.growth, than does the Greek. Hellenic outgrowths, of all kinds, began early to thrust forth themselves, in all the communities of Greece, with great force: so that, erelong, Hellenic, or civilized and cultivated, idea8, words, accent, and euphonism, with the power also of Hellenic arms, greatly changed the first character of the people and of their language. The Hellenes were, in a word, the Greeb, in a more cultivated period of their history, than the pioneer Pe­lasgi or first settlers. Thus readily is the great paradox solved, which has perplexed so long a succe8sion of historical writen!, of the connected existence, and yet supposed diverse origin,

o

1860.] Comparative Phonology. 301

of the two races, that peopled GreeC'.e, to which they really have themselves given, in their imagination, all the reality that t.hey ever possessed; and which it was as e81'1Y, of course, for them to set in grand ant.agonism to each other, as it was, in the first place, to invent them at all. Grecian literature, alt, history, genius, and advancement, are therefore, by necessity, all Hellenic.

The three leading dialects of Greece were the Aeolic, Do­ric, and Ionic. The AeoJic prevailed in Boeotia, The8ssly, and the colony of Aeolis in Asia Minor. The Doric, in the Peloponnesus and among t.he Dorian colonies in Asia Minor, Italy, amI Sicily; and the Ionic, which was spoken by the Ionian race, and especially in Asia Minor, and also in nume­rous islands, and in the Ionian colonies. This was the first of all the dialects, perfected by poetic composition; and it burst forth, at different times, into three kindred varieties: the Old Ionic or Epic, as seen in Homer and Hesiod, the New Ionic of Herodotus, and the Attic, which became ultimately the standard of all the other dialects, throughout the whole of Greece. This is the dialect, in which the many chief builders of Athenian greatness erected their various struoture8 of beaut.y and of strength.

The Aeolic and Doric are more tlimple, severe, and even rough, in their forms. In the .deolic, Alcaeus and Sappho sang. 'l'his is the dialect with which the forms of the Latin

• are more correlated than with any other; and which also presents to us the patterns of Greek words very nearly as they were at the first, when unbroken, or fused and recast into other moulds. The Doric abounds more in consonant&! than the ot.hers. In it the Muse of Theocritus and Pindar robed herself.

The Ionic is full of vowel~, and t.herefore soft; while, pos­sessing also an uncolltracted fulness of syllables, it moves before the eye like an Asiatic princess, with a Grecian face and smile, but sweeping a long train, and arrayed in the strong colors of the oriental world. The Attic dialect is the Ionic arrived at maturity. The hand of Time has here chiselled all its forms, according to the pure ideals of taste;

VOL. XVII. No. 66. 26

302 Comparative Phonology. [ APRIL,

the rules of art have been effectively applied to every side of it, by a long succession of workmen, busying them­selves one after another in perfecting the details of its structure; and it contains in itself all the plain, deep strength of the Doric, with all t.he real, upper beauty of the Ionic.

The Attic dialect had certain classes of peculiarities, in different ages, which have led writers sometimes, and yet with no very important results, to divide it into three periods: the Older, the Middle, and the Later Attic. The Older Attic flourished five hundred years before Christ, as found in the writings of Thucydides, Aeschylus, etc.; the Middle Attic, a hundred years later, as found in the works of Plato and Xenophon; and the Later Attic, in the succeeding age, as seen in the orations of Demosthenes. On the margin be­tween the Later Attic and the common Greek dialect, that prevai1ed 300 B. C., appeared that wonderful philosopher, Aristotle, who influenced the great speculative tjdes of thought. in the ancient world quite as much perhaps as Calvin has those of the modern. Some of the leading writers in the Common Greek, into which Classic Greek slowly, and with ever increasing dimness, faded away, were Plutarch, Strabo, Dionysius of Halicarnassus, and Lucian.

[To be continued.]

NOTB.- Errata and additions in the portion of this Article (Yo\. XVI.) pre· "ions\y published. Page 690, add after ras, in line 4, as part of the I!&me sen­tence: or a representation ofa+ i, as in amem forama·im. Page 69t,lines 30, 31, for dndtimi read dadhAmi. Page 693, line 18, for which read it. Page 7lS, line 2 from bottom, for going read agoing. l'age 722, transpose linea 3 and 4. Page 268,line 30, for form read from. Page 275, line 10, for become read became. Page 279, line 18, for when read where.

o